// Cam Z-02 · running

CV-2.1 · 30 fps · YOLO-v8

// Climate

19.4°C 92% RH CO₂ 820

// Detections this frame

147 leaves · 38 clusters · 1 anomaly

// Next harvest

Z-02 in ~38h · conf 0.91

AI Live The system is watching, right now

A grower that watches, listens, and explains itself.

AgriVision is the autonomous intelligence inside a small, modular vertical farm. Cameras read the crop frame by frame. The room runs itself. Nobody walks in to harvest a tray that wasn't ready.

Talk to us about your room → See what it sees

Scroll

01 Our
ethos

Three things we believe about growing food.

01 / 03

Small is strong.

A modest farm in a basement, run well by an autonomous system, beats a glass cathedral run on hope. We design for rooms, not warehouses.

02 / 03

Pull, not push.

Plant when there's a customer. Harvest when they're hungry. Nothing grown that doesn't have a plate waiting for it.

03 / 03

Explain everything.

Every decision the AI makes comes with a reason in plain language. The room runs itself, but never silently. You always know why.

02 Our
vision

Where we are going.

We are giving the farm a quiet intelligence. Ending the long journey of our food. Beginning the time when a city can feed itself, one small room at a time.

Most of what we eat starts in soil far away and ends on a truck. We think a different food system is possible — one where the room next door knows the crop better than any farm thousands of kilometers from a plate ever could.

Not a warehouse. Not a moonshot. A small, attentive farm in every neighborhood, run quietly by an AI that has learned what good growing looks like.

03 What the
camera sees

The room is empty. The model is looking.

Live · Z-02 · raw

AI · segmentation + detection

// Detection class

147 leaves

YOLO-v8 · conf ≥ 0.7

// Ripening

38 clusters

3 ready in <48h

// Anomalies

1 spotting

low priority · queued

// Inference

14ms/frame

on-device · Jetson Orin

Live · Z-05 · raw

AI · facility scan · cap counting

// Caps in room

799 tracked

L: 412 R: 387

// Ready to harvest

142 caps

Ø ≥ 7cm first flush

// Contamination

0.00%

no anomalies seen

// Climate · Z-05

18.2°C

94% RH · CO₂ 720

04 How it
works

Three layers of attention, working at three speeds.

Z-02 · 4K · vision-2.1

The same trays, every four hours. The model already knows what changed.

Reflex

The hands

Simple local rules running on small hardware. If humidity drops, the mister wakes up. If CO₂ climbs, the fan answers. The room takes care of itself, even if the network doesn't.

Speed · Always on

Sight

The eyes

A convolutional network watches the crop the way a careful grower would, only every few minutes. It learns what healthy looks like, what trouble looks like, and when something is ready to cut.

Cadence · Every few hours · on-device

Judgement

The mind

A larger model takes the long view: when to harvest, what to plant next, how to spend energy on a quiet night. Every call comes with its reasoning, written out — so the operator never has to wonder why.

Cadence · A few times a day

— autonomous monitoring · live feed

Z-01 · 06:24

cucumber bay · 147 leaves tracked

Z-02 · 14:11

LED bay · warm · canopy 0.96

Z-03 · 09:48

microgreens · day 11 · ready <48h

Z-04 · 11:02

tracking pass · clusters 38

Z-05 · 13:30

tray close-up · turgid · 0.91

Z-01 · 06:24

cucumber bay · 147 leaves tracked

Z-02 · 14:11

LED bay · warm · canopy 0.96

Z-03 · 09:48

microgreens · day 11 · ready <48h

Z-04 · 11:02

tracking pass · clusters 38

Z-05 · 13:30

tray close-up · turgid · 0.91

Live in production
Stockholm · 2026

Direct integration Mushu Mushrooms Lion's Mane, oyster, shiitake. 50 m² fruiting room. The AI watches every cluster, day and night.

Facility digitization Urban Seeds Shiso, amaranth, kresses, edible flowers. Research tent today, full data-driven facility next.

05 Who it's
for

Built for small rooms with serious intent.

Microgreens & herbs

Fast cycles, high care. The system loves tracking the small daily moves these crops make.

Leafy greens

Pak choi, kale, baby leaf. The kind of things people eat with dinner the same day they were cut.

Specialty mushrooms

Where the system was first proven. A demanding crop, watched closely from start to fruit.

Existing small farms

Already growing? AgriVision rides on top — same crop, same room, more attention paid.

— get in touch

Let's grow something quiet
and good.

If you run a vertical farm, a förening with an unused basement, or a kitchen that wants its herbs picked the same morning — write to us.

Send a note → See the system